Since 1998, two selective inhibitors of COX-2 have been approved in many countries for the treatment of rheumatoid arthritis, osteoarthritis and acute pain. These new drugs have a significantly reduced gastrointestinal toxicity when compared with non-selective COX inhibitors. However, the results of two large clinical trials conducted in patients with osteoarthritis and rheumatoid arthritis have recently raised some concerns regarding the cardiovascular safety of these new drugs. The purpose of this paper is to review the potential mechanisms whereby selective COX-2 inhibitors could increase the cardiovascular risk of patients and to analyse the data indicating that this clinical risk indeed exists. The authors' analysis shows that even though there are pathophysiological mechanisms which could explain why selective COX-2 inhibition might increase the cardiovascular risk in patients, the actual level of evidence demonstrating that the risk is indeed increased is weak. Because of the importance of the issue, additional studies must be conducted with this class of agents. Meanwhile, it is crucial to emphasise that neither selective COX-2 inhibitors nor conventional NSAIDs replace aspirin in patients with a high cardiovascular risk.